University of Groningen

Continuous metabolic monitoring techniquesTiessen, Renger Garmt

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2001

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Tiessen, R. G. (2001). Continuous metabolic monitoring techniques. s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-amendment.

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 08-01-2022

Chapter 2

Slow ultrafiltration for continuous in vivo sampling;application for glucose and lactate in man

Chapter 2

32

AbstractIntroduction: An ultrafiltration (UF) technique was developed for continuoussubcutaneous (s.c.) sampling and on-line analysis of absolute glucose andlactate concentrations in tissue. The relation between subcutaneous and bloodconcentrations was studied in men, because a subcutaneous monitoring devicewould put patients on less risks than an intravascular device.Methods: Ultrafiltrates were withdrawn continuously at a flow rate of 50-100nl/min from a hollow fibre probe to measure glucose in the abdominal subcutis.Six healthy volunteers underwent an oral glucose tolerance test. In order todetect glucose and lactate in the same sample, a splitter was placed between theon-line flow injection valve and the parallel enzymatic conversion andelectrochemical detection cells.Results: Subcutaneous glucose concentrations were in steady state on theaverage 1.06 mM lower. They rose delayed and blunted as compared to bloodlevels. We demonstrated the ability of simultaneous lactate and glucosemeasurements in vivo (n=2).Conclusions: UF makes continuous monitoring of absolute extracellularconcentrations in tissue possible. We interpret the deviations of subcutaneousmeasurements from intravascular levels in this way that the subcutis is a kineticcompartment not directly and exclusively linked to blood. The observeddifferences with blood suggest that diabetes management may demandintravascular monitoring. UF combined with analysis of glucose and lactate inthe same sample offers the opportunity to study pathophysiology inside tissues.

Glucose and lactate ultrafiltration___________________________________________________________

33

IntroductionIn clinical and basic research, pathophysiologic processes are often monitoredby in vivo sampling of body fluids. Sampling of blood is most commonlyutilised, because a large sample is relatively easy to obtain, and it can inform usabout all organs and there interactive functions. Clinicians increasingly relyingon fast biochemical analysis urge on continuous monitoring. However,continuous blood sampling poses infectious hazards and risks of internalbleedings, because the patient is often heparinised. Moreover, the wide tubesand high flow rates required to prevent obstruction can cause precious bloodloss to the patient. Finally, the concentration of the analyte may change rapidlyonce the blood is sampled and is no longer under physiological controlmechanisms.To avoid blood sampling, monitoring in vivo might done by biosensors. Sincetheir introduction in 1962(1), biosensors have been improved, thus creating awide variety of applications for off-line discontinuous monitoring in medicinee.g. for blood glucose control in medicine(2), for process control in industry(3),and for surface water control in the environment(4). Applications of biosensorsavailable for continuous in vivo monitoring of patients -bed-side or ambulant-are very limited. For such applications, the sensors should not only be safe,small, robust and easy to handle by any patient, but above all they must performreliably in the complex matrix of body fluids(5). Unfortunaley, sensors directlyintroduced in the body face bioincompatibility problems. Sampling bymicrodialysis (MD) and more recently by ultrafiltration (UF) has beenproposed as an interface between the body and the sensor(6-8), because thematerial used for both sampling methods has been extensively tested for safety.UF and MD use an implanted semi-permeable membrane to filtrate or dialysebiochemical analytes from the surrounding tissue interstitium. The perfusatewith the analytes is being transported by pumping and analysed outside thebody. Because the membrane excludes cells and large molecules, MD and UFare able to deliver a clean matrix for measurement to a biosensor. Anotheradvantage is the ability to monitor processes inside tissue of specific organs(6).In this way, the afore-mentioned intravascular hazards are evaded.In order to use subcutaneous MD for clinical decision making, thesemeasurements should invariably be closely related to blood. A major problemof MD is the exact determination of the recovery of the analyte in vivo becauseit is lower and more variable than in vitro. The (relative) recovery is defined asthe dialysate/body-fluid concentration ratio expressed as a percentage. Therecovery

Chapter 2

34

is assumed to be independent of the concentration of the analyte(9). Sometimes,the absolute recovery is used, being the amount of analyte recovered in thedialysate during a certain time interval. Despite the development of severaltechniques such as ‘no-net-flux’(9), ‘internal reference’(10), and ‘various flowrates(11)’ to determine the recovery in vivo, it remains difficult to estimate theabsolute concentrations with MD. To follow only trends with relativemeasurements limits the applicability of the MD technique for monitoringpatients. MD has some other drawbacks, e.g. the recovery may change afterlong-term implantation (possibly due to adhesion of fibrin, collagen, and cellsafter traumatic probe insertion), and there may be a removal of the analyte orinflux of MD buffer which is not rapidly compensated. So, MD may alsoinfluence the physiology at the sampling site. An important question about therelation of subcutaneous MD to blood remains yet unanswered. Why are MDconcentrations lower and more variable between probes in vivo as compared toin vitro? Are these differences entirely due to changes of the probe, or are thesedifferences (also) reflecting actual lower tissue concentrations at the site ofmeasurement? The latter hypothesis carries the consequence that glucoseconcentrations in the subcutis interstitium would be not merely dependent onintravascular concentrations but would be also dependent on local factors of theprobe surrounding tissue. If that would be the case, subcutaneous glucose levelswould be the result of the local equilibrium between supply from the vessels(influx) and uptake by cells (efflux). As the subcutis itself produces no glucoseto buffer changes, the subcutaneous glucose levels might be more variable thanin blood. In fig.1. we show four types of subcutaneous glucose curves to beexpected in an OGTT. In case the glucose efflux is large compared to theinflux, the glucose levels will be lower (fig.1.C,D) than when the influx is moredominant (fig.1 A,B). The equilibrium between influx and efflux may settle fastor slow, resulting in a steep follow-up curve (fig.1.A,C) or a flattened curve(fig.1.B,D). To test the hypothesis of lower glucose concentrations in tissue, asubcutaneous detection method is needed which combines measurement ofabsolute concentrations with a high time resolution. UF is an alternativesampling technique for MD, developed to make the analyte concentrationindependent from diffusion through the probe membrane. The UF techniquesamples interstitial fluid through a membrane by underpressure instead ofdiffusion. The semi-permeable membrane used for UF sampling is comparablewith the microdialysis membranes, and excludes large molecules (e.g. largeproteins), whereas small analytes such as glucose and lactate enter the probetogether with water and ions. The technique was firstly described for batch-wise sampling(12;13) with large

Glucose and lactate ultrafiltration___________________________________________________________

35

Fig. 1. four types of subcutaneous glucose curves to be expected in anOGTT. Left-hand side: schematic representation of different glucose influx(from capillaries) and efflux (to cells) from the s.c. interstitium. Right-handside: resulting s.c. concentration curves.A,B: glucose efflux is small compared to glucose influx;the s.c.glucose levels will be close to i.v. levelsC,D: glucose efflux is large compared to glucose influx;the s.c.glucose levels will be lower than i.v. levelsA,C: the equilibrium between influx and efflux settles fast,resulting in a steep follow-up curveB,D: the equilibrium between influx and efflux settles slow,resulting in a flattened follow-up curve

capillary cell top line: intravascular glucosebottom line: expected subcutaneous glucose

A

B

C

D

subcutaneousinterstitium

Chapter 2

36

probes and a low sampling frequency (every 10-20min). Recently, small UFprobes have been proposed(14;15). In our laboratory Moscone et al.(15) andKaptein et al.(16) developed an UF technique with a 4 cm probe for continuouson-line sampling and analysis. A disposable pump (weight 5 g) produces astable 100-300 nl/min sample flow rate for 24 hours and longer. Unlike MD,UF sample concentrations are not dependent on probe diffusion characteristicsbecause analytes of low molecular size are almost 100% recovered by solventdrag(16;17). Large molecules, however, may encounter hinderance of themembrane, leading to an underestimation of the analyte concentration.Discontinuous subcutaneous UF-sampling technique with large probes has beendocumented well(7;12;13;17;18). The method developed in our laboratory uses smallprobes and allows frequent analysis(15;16).Here, the potential of the on-line ultrafiltration technique in vivo in humanvolunteers was explored to test the hypothesis of lower glucose concentrationsin tissue. UF fluid was continuously withdrawn from the extracellular space inhuman subcutis. The UF fluid was analysed discontinuously every one or twominutes by a flow injection system detecting glucose electrochemically afterenzymatic conversions. In this manner, glucose levels were continuously moni-tored in six volunteers who underwent an Oral Glucose Tolerance Test(OGTT). As stated above, the hypothesised lower s.c. glucose levels imply theexistance of an interaction between influx and efflux parameters. Theseparameters would influence the levels of other metabolites in tissue as well. Wetested this for lactate. The system was improved to analyse glucose and lactatelevels in the same sample. Two subjects were monitored in this manner.

MethodsGeneral description of the experimentsThe experiments were performed in six volunteers. The UF probes were placedsubcutaneously (sc) near the umbilicus. During the experiments, the UF-probeswere connected to the glucose and lactate detection system. The ultrafiltrationsampling and analysis set-up is shown in figure 2. The sampling part on thebottom side comprised a hollow fibre probe and a semi-vacuum tube as thedriving force on a balance. The sample was brought into the detection part ofthe system by an intercalated switching loop. The detection system(15;19)

consisted of a Phosphate Buffered Saline (PBS) solution containing ferrocene,an HPLC pump, a streamsplitter, two parallel enzyme reactors containing HorseRadish Peroxidase (HRP) and either Lactate or Glucose Oxidase (LO or GOD),and two parallel ElectroChemical cells (ECDs). The ultrafiltrate was analysed

Glucose and lactate ultrafiltration___________________________________________________________

37

every one or two minutes by switching the position of the valve, which injectedultrafiltrate into the detection system.

Fig. 2. Ultrafiltration sampling and analysis set-up. bottom side: samplinghollow fiber probe and UF pump on a balance; middle: sample injection valve(loop in injection position___ or in load position---); top side: detection withPBS buffer containing ferrocene, HPLC pump, splitter, two parallel enzymereactors containing Horse Radish Peroxidase (HRP) and either Lactate orGlucose Oxidase (LO or GOD), and two parallel Electric detectors (ECDs).

UltrafiltrationThe probe consisted of a 2.5 cm long hollow fibre from an artificial kidney(AN69HF; Hospal Ind., Meyzieu, France, 340µm OD, 240µm ID, MWCO50kD). To keep the lumen patent, the fiber was reenforced from the inside witha home-made spring of stainless steel wire(Vogelsang, Hagen, Germany),D=60µm, 12 axial length windings/cm. The fiber was glued with cyanoacrylate(Sicomet 40; Henkel Corp., Kankakee, IN, USA) to a draining fused silicacapillary tube (20-65cm l, 150µm OD, 50µm ID, Polymicro Technologies inc.,Phoenix, AZ, USA). Before closing the other end of the finer with glue, thetube and finer were filled up with 0.9% NaCl. The probe was placed in 0.9%NaCl in a disposable syringe before sterilisation with gamma irradiation (min.25 kiloGray; Gammaster B.V., Ede, the Netherlands). The ultrafiltration flow rate was generated by the semi-vacuum (625±1mbar±SD, 0.2 %CV) of a syringe (1.2 ml monovette, Sarstedt, Nümbrecht,Germany) with a fixed piston(19). To stabilise the ultrafiltration flow a definedcapillary restriction (l=4cm, OD=150µm, ID=15µm fused silica tube;Polymicro Technologies inc., Phoenix, AZ, USA) was placed in front of thesemivacuum. To avoid flow disturbance a fluid-filled bubbletrap was placedbetween the probe draining tube and the capillary restriction. The time

WastePBS withferrocene

ECDsHPLCPump

Valve

Loop

UF pumpProbe

Waste

god + hrpreactors

lo + hrp

Balance

Chapter 2

38

resolution of the system is limited by spreading of 3-5 minutes(16).The in vivo ultrafiltration flow rate was measured by weight. The subcutaneousflow rate was on average 47 nl/min (range 32-100 nl/min), as compared to 100nl/min in vitro. When the ultrafiltrate flow changed, the instrumental lag-timeand measurements were corrected accordingly.

Combined glucose and lactate analysisThe flow injection analysis system described by Elekes et al.(19) was modifiedto detect glucose and lactate in the same sample. A splitter, a Lactate Oxidasewith Horse Radish Peroxidase enzymatic cell, and an Electrochemical cell wereadded to the set-up (see fig.2). The UF sampling part was connected to theanalytical part of the set-up by an intercalated valve. A Decade sampler (AntecLeyden B.V., Zoetermeer, The Netherlands) equipped with a Vici CheminertC4 valve with a 20nl internal loop (Valco Instruments Co. Inc., Houston, USA)was used for sample injection in the detection system. The loop was partiallyfilled. The valve injected every 15 seconds a 2 seconds collected ultrafiltratesample in ferrocene PBS, which was being pumped at a flow rate of 0.8 ml/min(HPLC pump LC-10AD, Shimadzu, Japan) through a fifty-fifty splitter and twoenzyme reactors and electrochemic detectors. The flow split proportion wasapproximately 50/50 and remained stable throughout the experiments.Ferrocene PBS contents were: 137 mM NaCl, 2.7mM KCl, 8mM Na2HPO4,2.5 mM KH2PO4 (pro-analysis quality purchased from Merck, Darmstadt,Germany),0.5 mM Ferrocenemonocarboxylic acid (Sigma Chemical Co, St.Louis, MO, USA) and 0.1 volume% Kathon CG (Rhom and Haas, Croydon,UK) in double quarts distilled water. The solution was neutralised to pH 7.4,filtered, and bubbled with Helium to remove air. The enzyme reactorscontained 200 U Horseradish Peroxidase (EC 1.11.1.7) and either 200 UGlucose Oxidase (EC 1.1.3.4, grade I) or 200 U Lactate Oxidase Lyophilizate(Boehringer Mannheim, Germany) in 20 µl 0.9% NaCl immobilised betweencellulose nitrate filters (pore size 0.01µm, MWCO 50kD, Sartorius, Göttingen,Germany). The electrochemical cells were of a thin layer-type, with a workingelectrode of glassy carbon kept at 0.00 mV relative to an Ag/AgCl referenceelectrode and a teflon/carbon counter electrode (Amor cell, Spark Holland,Emmen, The Netherlands). The potentiostates were a Decade (Antec LeydenB.V., Zoetermeer, The Netherlands) and an Amor amperometric detector(Spark Holland, Emmen, The Netherlands).Calibration curves were made with solutions containing 0, 1, 2, 4, 6 mM lactateand 0, 2, 4, 8, 12 mM glucose, changed stepwise every 5 minutes, increasingfrom 0 mM and decreasing with the same steps. The spreading in the analytical

Glucose and lactate ultrafiltration___________________________________________________________

39

system due to instrumental mixture between consecutive samples was definedas the time between a 20 % and 80% amperometric signal change betweenconcentrations. The spreading ± SEM was calculated applying sigmoidal-fittingto the steps in the calibration curve.

Oral Glucose Tolerance Test (OGTT)Six OGTTs were performed with healthy young males and females of normalweight. Following an overnight fast, subjects sat in an easy chair from 8 a.m.till 1 p.m. At zero time they drank 100 g glucose dissolved in 200 ml tea. Anultrafiltration probe was placed subcutaneously near the umbilicus using a 22 Gcannula at approximately 45 minutes before the start of the OGTT. Glucose inultrafiltrates was detected electrochemically using a bi-enzyme reactor in aflow injection system as described extensively by Elekes et al.(19).An ECA 180 Glucoanalyser (Medingen GmbH, Dresden, Germany) measuredblood samples taken from a forearm vein cannula every 5 to 15 minutes.Glucose molarity (mol/L blood) of haemolysed whole blood is different fromthe molality (mol/kg water) as about 80% of the blood volume consists ofwater(20). The molality of glucose is the same in whole blood, bloodplasma, andthe extracellular fluid in blood. To make a correct comparison with the UFconcentrations, whole blood concentrations were increased with 17.6% inaccordance to the manufactorer's report. Steady state levels were calculated asthe average of the first three i.v. samples (t=-15,0,5 min) and the concurrent s.c.levels. Lactate was determined in blood plasma with the Vitros 250 (OrthoClinical Diagnostics, Beerse, Belgium). No correction for plasma proteinvolume was done, because no corrective factor was available.

Fig. 3. Amperometric recording of calibration. Upper part: lactate. Lower part:glucose. Standard concentrations of glucose and lactate were changed every 5minutes (↑ = artefact).

Chapter 2

40

ResultsIn vitro experimentsAt the applied flow rate of 100nl/min, 3.3 nl was injected 4 times per minuteinto the analytical system. A photograph of an amperometric recording of acalibration curve is shown in fig.3. Standard concentrations of glucose andlactate were changed every 5 minutes. The respons was 4 nA/mM lactate and0.3 nA/mM glucose. Regression analyses for the lactate and glucoseconcentrations with amperometric top-baseline values showed linearity for 0-6mM lactate and 0-12 mM glucose (both r > 0.99, p < 0.0001). The spreading inthe analytical system was 32±5 seconds. An artefact was produced throughsample flow disturbance by roughly changing test concentrations (↑ = artefact).

Oral Glucose Tolerance Test (OGTT)In the steady state before the OGTT the glucose concentration was on average1.06 mM lower subcutaneously than intravenously (95% c.i. 0.127-1.98 in apaired t-test). We found no correlation between the ultrafiltrate flow rate andthese steady state levels (r = 0.14). Figure 4 shows the results of the OGTT in

0

2

4

6

8

10

12

23:30 0:00 0:30 1:00 1:30 2:00 2:30

0

2

4

6

8

10

12

23:30

0:00 0:30 1:00 1:30 2:00 2:30

Glu

cose

(mM

)

0

2

4

6

8

10

12

23:30

0:00 0:30 1:00 1:30 2:00 2:30

Tim e (hours)

Glu

cose

(mM

)

0

2

4

6

8

10

12

23:30 0:00 0:30 1:00 1:30 2:00 2:30

Tim e (hours)

A B

C D

Fig. 4. OGTT (|) at zero time in four healthy volunteers. glucose levels in bloodsamples (◆ ) and in subcutaneous ultrafiltrates (�)

Glucose and lactate ultrafiltration___________________________________________________________

41

four healthy volunteers (A/D). After glucose ingestion at time zero, thesubcutaneous levels follow the intravenous glucose rise with various speed. Thetimecourse subcutaneous was not only delayed, but also less steep thanintravenous. Delay times between the i.v and s.c. maximum glucose levelranged from about one minute in C to 30 minutes in B.Because we observed systematic differences between i.v. and s.c. levels, wethought it not appropriate to lump these data to perform regression analysis.

Combined glucose and lactate analysis in vivoFig.5 shows the results of two experiments (A,B and C,D) in which wemeasured both glucose and lactate. In one experiment, the subcutaneous lactatelevels equalled blood levels where glucose levels did too (A and C). In thesecond experiment, glucose were lower and flattened compared to intravenous(B). Lactate was higher subcutaneous (D). So the s.c.-i.v. relation for bothglucose and lactate appeared to be close in one experiment, whereas it appeareddistant in the other experiment.

Fig. 5. OGTT (|) at zero time in two healthy volunteers (A,C and B,D). Simultaneousmeasurement of glucose and lactate levels in blood samples (◆ ) and in subcutaneousultrafiltrates (�)

0

2

4

6

8

10

12

23:30 0:00 0:30 1:00 1:30 2:00 2:30

Glu

cose

(mM

)

012345678

23:30 0:00 0:30 1:00 1:30 2:00 2:30

Tim e (hours )

012345678

23:30 0:00 0:30 1:00 1:30 2:00 2:30

Time (hours)

Lact

ate

(mM

)

0

2

4

6

8

10

12

23:30 0:00 0:30 1:00 1:30 2:00 2:30

A B

C D

Chapter 2

42

ConclusionsThe present study was aimed to test the hypothesis that s.c.glucose levels arelower than i.v. levels. Two of our previously developed techniques werecombined, i.e. a slow UF technique(15) to obtain absolute concentrations, and aflow injection analysis of glucose(19) to obtain a high time resolution. The flowinjection system was further expanded in order to determine lactate and glucosein the same sample.The simultaneous in vitro glucose and lactate analysis presented, has severaladvantages over previously used techniques. The linearity of the detection iselongated by dilution through injection of very small samples. The pulse-freesampling pump makes it possible to lower the sample volume, and to increasethe injection frequency without pulse related problems. We created newresearch opportunities by introducing a splitter in the flow injection system,thus realising analysis of two metabolites in one sample. Having the advantage of a system for frequent analysis of absoluteconcentrations, UF research was done in the human subcutis. Some cleardifferences were found between blood and subcutaneous glucoseconcentrations in an OGTT in men. The ultrafiltration drainage rate wascompared with glucose levels in steady state and these did not show anycorrelation. Moreover, the fluid volume being removed equals the blood flowthrough 2 mm3 adipose tissue (assuming 0.03 ml/ml/min adipose tissue bloodflow). This leads us to believe that UF disturbs local tissue physiology onlylittle. The differences found in these experiments between subcutaneousglucose levels and intravenous levels suggest that these are not simply linearlyrelated. Rather, various lower steady state concentrations, time delay, and lesssteep changes in the subcutis were found in these first experiments (figs.4 and5A,B). The measured curves fit well in the expected curves (fig.1) and thussupport the presented hypothesis on lower s.c. glucose concentrations. Thepreliminary results obtained by lactate and glucose measurements in the samesample enable to interpret both measurements together. In the first case s.c.levels of both glucose and lactate are in close relation to the i.v. levels (fig.5A,C), compatible with curve A in fig.1. In the second case is the glucose curve(fig.5 B) compatible with curve type D in fig.1. Curve D is characterised by ahigh efflux of glucose into cells compared to the influx of glucose from thecapillaries. Lactate concentration is higher than intravenous in this condition; asto be expected, because it is a metabolite of glucose traveling the reverse way.Previous UF-experiments in rats(16) showed similar results as in men. Thedifferences in rats between s.c. and i.v. glucose concentrations were recentlyconfirmed(21). Others reported human steady state s.c. concentrations in the

Glucose and lactate ultrafiltration___________________________________________________________

43

same range as found here; Schmidt et al reported subcutaneously 44±8% and46±9% of blood concentration(22), Sternberg et al 72±6%(23), Stallknecht85%(24), Lönnroth et al 91±9 %(10), and Bolinder et al reported 91±2%(25). Thesedata confirm our conclusion that the s.c. glucose concentration is not directlyand exclusively linked to blood. Further research may yield the various factorswhich modify the glucose influx and efflux, and eventually a s.c. kinetic model.Detection of both lactate and glucose in the same s.c. space may be useful toanalyse local tissue metabolism. The relationship between i.v. and s.c. glucosehowever, seems far from direct as has been assumed previously(25). So, theextracellular space of abdominal fat tissue seems less suitable for sampling andcontrol of glucose in diabetes patients, than has often been proposed(5).Mascini(26) tested a MD probe in an other location, the forearm subcutis, withsimilar results results as in our abdominal experiments. Thus, a different s.c.location might not resolve the problem. We feel that to avoid complicatedkinetic models, glucose sensors need to be placed intravascularly.The present study illustrates the potency of slow continuous UF sampling insubcutaneous tissue. The combination of biosensor technology and continuousin vivo sampling with UF may lead to the development of biochemicalmonitoring devices. Such devices will contribute to clinical and basic researche.g. the study of (patho-)physiology of energy metabolism in particular tissues.

References

1. Clark, L. and C. Lyons. Electrode systems for continuous monitoringcardiovascular surgery. Ann N Y Acad Sci 102: 29-45, 1962.

2. Pickup, J.C. Sampling and sensing blood glucose [comment]. Lancet 342:10681993.

3. Renneberg, R., G. TrottKriegeskorte, M. Lietz, V. Jager, M. Pawlowa, G.Kaiser, U. Wollenberger, F. Schubert, R. Wagner, and R.D. Schmid. J.Biotechnol. 21: 1731991.

4. Sadik, O.A. and J.M.v. Emon. Biosens.Bioelectron. 11: i-ix1996.5. Reach, G. and G.S. Wilson. Anal.Chem. 64: 381A1992.6. Boer, J.d., J. Korf, and G.H. Plijter. In vivo monitoring of lactate and glucose

with microdialysis and enzyme reactors in intensive care medicine. Int.J.Artif.Organs 17: 163-170, 1994.

7. Ash, S.R., J.T. Poulos, J.B. Rainier, W.E. Zopp, E. Janle, and P.T. Kissinger.Subcutaneous capillary filtrate collector for measurement of blood glucose. ASAIO.J. 38: M416-M4201992.

8. Rosdahl, H., Ungerstedt U., L. Jorfeldt, and J. Hendrikson. J.Physiol.Lond.471: 6371993.

9. Lonnroth P, P.A. Jansson, and Smith U. A microdialysis method allowing

Chapter 2

44

characterization of intercellular water space in humans. Am.J.Physiol. 253:E228-E2311987.

10. Lonnroth, P. and L. Strindberg. Validation of the 'internal reference technique'for calibrating microdialysis catheters in situ. Acta Physiol.Scand. 153: 375-380,1995.

11. Boer, J.d., F. Postema, G.H. Plijter, and J. Korf. Continuous monitoring ofextracellular lactate concentration by microdialysis lactography for the study ofrat muscle metabolism in vivo. Pflugers Arch. 419: 1-6, 1991.

12. Ash, S.R., J.B. Rainier, W.E. Zopp, R.B. Truitt, E.M. Janle, P.T. Kissinger,and J.T. Poulos. A subcutaneous capillary filtrate collector for measurement ofblood chemistries. ASAIO.J. 39: M699-M7051993.

13. Linhares, M.C. and P.T. Kissinger. J.Pharm.Biomed.Anal. 11: 11211993.14. Schneiderheinze, J.M. and B.L. Hogan. Anal.Chem. 68: 37581996.15. Moscone, D., K. Venema, and J. Korf. Ultrafiltrate sampling device for

continuous monitoring. Med.Biol.Eng.Comput. 34: 290-294, 1996.16. Kaptein, W.A., R.H. Kemper, M.H. Ruiters, K. Venema, R.G. Tiessen, and

J. Korf. Methodological aspects of glucose monitoring with a slow continuoussubcutaneous and intravenous ultrafiltration system in rats. Biosens.Bioelectron.12: 967-976, 1997.

17. Linhares, M.C. and P.T. Kissinger. Anal.Chem. 64: 28311992.18. Linhares, M.C. and P.T. Kissinger. Pharm.Res. 10: 5981993.19. Elekes, O., D. Moscone, K. Venema, and J. Korf. Bi-enzyme reactor for

electrochemical detection of low concentrations of uric acid and glucose. Clin.Chim.Acta 239: 153-165, 1995.

20. Marks, V. Blood glucose: its measurement and clinical importance. Clin.Chim.Acta 251: 3-17, 1996.

21. Schmidtke, D.W., A.C. Freeland, A. Heller, and R.T. Bonnecaze.Measurement and modeling of the transient difference between blood andsubcutaneous glucose concentrations in the rat after injection of insulin. Proc.Natl.Acad.Sci.U.S.A. 95: 294-299, 1998.

22. Schmidt, F.J., W.J. Sluiter, and A.J. Schoonen. Glucose concentration insubcutaneous extracellular space [see comments]. Diabetes Care 16: 695-700,1993.

23. Sternberg, F., C. Meyerhoff, F.J. Mennel, F. Bischof, and E.F. Pfeiffer.Subcutaneous glucose concentration in humans. Real estimation and continuousmonitoring. Diabetes Care 18: 1266-1269, 1995.

24. Stallknecht B. Interstitial concentrations of glucose, lactate and glycerol. Estimation of interstitial concentration of metabolites and hormones in adiposetissue by means of microdialysis - method evaluation (thesis) 40-42, 1996.

25. Bolinder, J., U. Ungerstedt, and P. Arner. Long-term continuous glucosemonitoring with microdialysis in ambulatory insulin-dependent diabetic patients[see comments]. Lancet 342: 1080-1085, 1993.

26. Mascini, M. Biomed.Eng. 16: 931996.